[英]How to make a redis cluster in k8s environment using nodeport service type?
我嘗試使用“NodePort”類型的服務在 k8s 環境中創建 redis 集群。 更具體地說,我想跨兩個不同的 k8s 集群組成一個 redis 集群。
當我使用 LoadBalancer(External IP) 作為服務類型時,集群已成功創建。 問題是節點端口。
在我命令redis-cli --cluster create之后,它停留在“等待集群加入”
以下是集群創建命令的日志。 我使用單獨的 nodeport 服務部署了 4 個領導 pod 和 4 個從屬 pod。
root@redis-leader00-5fc546c4bd-28x8w:/data# redis-cli -a mypassword --cluster create --cluster-replicas 1 \
> 192.168.9.194:30030 192.168.9.199:30031 192.168.9.194:30032 192.168.9.199:30033 \
> 192.168.9.199:30030 192.168.9.194:30031 192.168.9.199:30032 192.168.9.194:30033
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
>>> Performing hash slots allocation on 8 nodes...
Master[0] -> Slots 0 - 4095
Master[1] -> Slots 4096 - 8191
Master[2] -> Slots 8192 - 12287
Master[3] -> Slots 12288 - 16383
Adding replica 192.168.9.199:30030 to 192.168.9.194:30030
Adding replica 192.168.9.194:30033 to 192.168.9.199:30031
Adding replica 192.168.9.199:30032 to 192.168.9.194:30032
Adding replica 192.168.9.194:30031 to 192.168.9.199:30033
M: 94bf3c6760e6b3a91c408eda497822b4961e8d82 192.168.9.194:30030
slots:[0-4095] (4096 slots) master
M: 31f4a9604b15109316f91956aa4a32b0c6952a4d 192.168.9.199:30031
slots:[4096-8191] (4096 slots) master
M: 0738d1e1a677352fc3b0b3600a67d837b795fa8a 192.168.9.194:30032
slots:[8192-12287] (4096 slots) master
M: 7dd7edbfab6952273460778d1f140b0716118042 192.168.9.199:30033
slots:[12288-16383] (4096 slots) master
S: 17e044681319d0a05bd92deeb4ead31c0cd468e2 192.168.9.199:30030
replicates 94bf3c6760e6b3a91c408eda497822b4961e8d82
S: 9c9e47ec566ac781e8e3dcb51398a27d1da71004 192.168.9.194:30031
replicates 7dd7edbfab6952273460778d1f140b0716118042
S: b8f7028b56f96565a91fdb442c94fbedcee088c2 192.168.9.199:30032
replicates 0738d1e1a677352fc3b0b3600a67d837b795fa8a
S: e4c9ffdf67e2b2ef9f840618110738358bde52d5 192.168.9.194:30033
replicates 31f4a9604b15109316f91956aa4a32b0c6952a4d
Can I set the above configuration? (type 'yes' to accept): yes
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join
...........................................
奇怪的是,其他 redis 容器收到了信號但沒有任何進展。 以下是其他 redis 容器的日志。
.... // Some logs for initializing redis
1:M 20 Jan 2022 06:09:12.055 * Ready to accept connections
1:M 20 Jan 2022 06:13:41.263 # configEpoch set to 5 via CLUSTER SET-CONFIG-EPOCH
我以為通信成功了,但是 gossip 端口無法正常工作。 所以,我修改了 redis.conf 並設置了 cluster-announce-bus-port 但它也沒有工作。
如何使用節點端口類型的服務組成 redis 集群?
請參考.yaml文件之一
apiVersion: v1
kind: ConfigMap
metadata:
name: redis-cluster-leader00
namespace: redis
labels:
app: redis-cluster
leader: "00"
data:
fix-ip.sh: |
#!/bin/sh
CLUSTER_CONFIG="/data/nodes.conf"
if [ -f ${CLUSTER_CONFIG} ]; then
if [ -z "${HOST_IP}" ]; then
echo "Unable to determine Pod IP address!"
exit 1
fi
echo "Updating my IP to ${HOST_IP} in ${CLUSTER_CONFIG}"
sed -i.bak -e "/myself/ s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/${HOST_IP}/" ${CLUSTER_CONFIG}
fi
exec "$@"
redis.conf: |+
bind 0.0.0.0
cluster-enabled yes
cluster-require-full-coverage no
cluster-node-timeout 15000
cluster-config-file /data/nodes.conf
cluster-migration-barrier 1
appendonly no
save ""
protected-mode no
requirepass "mypassword"
masterauth "mypassword"
cluster-announce-ip 192.168.9.194
cluster-announce-port 30030
cluster-announce-bus-port 31030
---
apiVersion: v1
kind: Service
metadata:
name: redis-leader00
namespace: redis
labels:
app: redis
role: leader
tier: backend
leader: "00"
spec:
ports:
- port: 6379
targetPort: 6379
nodePort: 30030
name: client
- port: 16379
targetPort: 16379
nodePort: 31030
name: gossip
selector:
app: redis
role: leader
tier: backend
leader: "00"
type: NodePort
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: redis-leader00
namespace: redis
labels:
app: redis
role: leader
tier: backend
leader: "00"
spec:
replicas: 1
selector:
matchLabels:
app: redis
leader: "00"
template:
metadata:
labels:
app: redis
role: leader
tier: backend
leader: "00"
spec:
containers:
- name: leader
image: docker.io/redis:6.0.5
resources:
requests:
cpu: 100m
memory: 100Mi
ports:
- containerPort: 6379
name: client
- containerPort: 16379
name: gossip
volumeMounts:
- name: conf
mountPath: /conf
readOnly: false
args: ["--requirepass", "mypassword"]
command: ["/conf/fix-ip.sh", "redis-server", "/conf/redis.conf"]
env:
- name: HOST_IP
valueFrom:
fieldRef:
fieldPath: status.hostIP
volumes:
- name: conf
configMap:
name: redis-cluster-leader00
defaultMode: 0755
在我命令redis-cli create后,還可以查看容器中的 nodes.conf 文件
root@redis-leader01-87ccb466-bsnq4:/data# cat nodes.conf
31f4a9604b15109316f91956aa4a32b0c6952a4d 192.168.9.199:30031@31031 myself,master - 0 0 2 connected 4096-8191
vars currentEpoch 2 lastVoteEpoch 0
我不確定您創建 Redis 集群的實際過程,但是我建議您查看 helm 圖表以在 K8s 上部署 Redis 集群。
使用 helm chart 可以輕松地在 K8s 上管理和部署 Redis 集群。
https://github.com/bitnami/charts/tree/master/bitnami/redis
要部署圖表,您只需運行命令:
helm install my-release bitnami/redis
在節點端口側,一旦部署了 helm chart,您可以編輯服務類型,否則您還可以更新第一個 helm chart,然后將這些更改應用到 K8s。
這將在 K8s 上為 Redis 服務創建節點端口。
你可以在這里找到服務模板: https://github.com/bitnami/charts/tree/master/bitnami/redis/templates
還有其他人已經發現了問題。 這個鏈接是他提出的問題。 https://github.com/redis/redis/issues/6432
問題的原因是 CLUSTER MEET 沒有使用 cluster-announce-bus-port。 實際上,在 clusterManagerNode 結構中甚至沒有用於存儲 cluster-announce-bus-port 的變量。
這個問題已經有 PR 了。 https://github.com/redis/redis/pull/6442
它尚未合並或關閉,所以這是持續存在的問題......
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.